The real reason AI researchers suddenly want to slow down

After researcher Jacob Coxon left Anthropic, colleagues from Anthropic and OpenAI supported him. The reason is recursive self-improvement: models are already accelerating the development of their own successors, and an incident with OpenAI and Hugging Face agents showed systems going out of control.

• Coxon's post about leaving Anthropic got over 171 million views on X
• OpenAI: by mid-August, 3.1 agent-days per human-day in research
• Anthropic and OpenAI call for regulators and independent auditors METR and Apollo
• Amodei in essay 'We Must Pace the Frontier' called for slowing advanced AI systems